CD Laboratory for Mastering Variability in Software-intensive Cyber-physical Production Systems

Production systems such as metallurgical plants and car manufacturing plants are highly complex individual solutions. Constantly changing requirements pose major challenges for industry. This CD Laboratory attempts to model the variability of such systems in order to increase the degree of reuse, adaptability and configurability.

A cyber-physical system (CPS) is a highly complex, distributed system in which physical components and software components are closely networked and interact with each other. CPS in production, such as metallurgical plants and car manufacturing plants, are called (software-intensive) cyber-physical production systems (SiCPPS). SiCPPS, are ubiquitous systems that autonomously interact with their environments and are capable of flexibly manufacturing a variety of products based on customer demands.

Typically, such systems comprise a large number of heterogeneous components (mechanical, electrical, mechatronic, software) that can be configured and combined in different ways. They are highly variable and are subject to constant evolution during their typically long lifetime.

A main driver of variability is the hardware, but also development processes, application domains, methods and tools. Another major influencing factor of variability results from market pressure and customer requirements.

Industry is therefore keen to master variability as it is a prerequisite for successful reuse, which in turn helps to reduce development, certification and maintenance costs and shortens time to market.

Dealing with variability currently depends too much on domain and expert knowledge, as well as on self-developed tools. These often only work for certain hardware and software platforms. Industry has suffered for many years from a lack of methods and tools that support it in rapidly adapting its systems to constantly changing requirements.

A main focus of this CD Laboratory will be on automatically handling variability, e.g., analyzing existing SiCPPS artifacts from the design process of SiCPPS, to automatically extract and model variability information and on generating and configuring target artifacts. This will allow companies developing SiCPPS to better support system variability, system evolution as well as future changes in software platforms, hardware platforms, and tools.